Yeah i fear that too, my grandma is already sending me links of AI animals that she thinks is real, and the horrible/beautiful art of facebook memes/holiday cards, seems to be completely overtaken by AI. We know that full fake video of you with your own voice asking for something or even interacting on a video call is basically solved problem. Prime time to reestablish and confirm trusted channels with the people you care about.

Recently at a family dinner we established that any kind of unsolicited contact that falls outside typical conversation - Asking for money, sending money, pretty much anything with money - you must say the word that only people in our family would know.

Ironically, this would be a good application of AI, where the AI listens in on their calls, and will flag conversation that warrants the keyword being said.

Do you think anybody will really care? People were generating CSAM basically as soon as image generation become accessible. And for the less dangerous stuff situation is way more rampant already, both in free and commercial way.

Yes. Deepfaked porn is already a widespread mechanism for harassment, both among children and adults. As it gets even easier to create and more and more convincing it will just get worse.

Different way people learn, and different way they create. Other creative fields have that dichotomy too, but i guess we are closer to math, and always try to find "correct way" of doing things, while our field is actually really flexible and allowing for a lot of self expression.

I recently started drawing (my own little middle age crisis) and i am so out of depth, so far from even understanding what is possible, and the internet filled with so much junk tutorials, that sometimes ability of LLM to parse my intent and guide me towards the key words and names is a godsend, of course it probably would be better to ask a person who knows this stuff, but it's not always convenient. In my use case best part of LLMs is Soft input, and the soft output is not a problem.

For drawing it’s usually better to buy a book or a course (from a good platform like New Masters Academy), and then practice (a lot)

This is not a LLM feature, but a failure of traditional search engines.

In my experience hauling deliveries on the shoulders generally sucks, especially if weight is bigger. hence why most active couriers around my parts prefer to fix some sort of baggage rack to their ebikes.

Capacity is still needed, or you will miss out on a lot of deliveries in peak hours, where you load up several orders and then deliver them all in one route and you dont really need to carry whole bag to client/restaurant, just fish up the package. Also boxy bags allow for bigger packages(pizza, water bottles) and usually help to pack orders in a way that damages them less.

Which companies batch food orders now? It was a short lived and unpopular experiment in India years back.

Both Grubhub and DoorDash allow drivers to pickup multiple orders in one run in my neighborhood.

Well it is pretty normal in russia, atleast for big operators

Can the claim about citation frequency be just an answer pattern and not model's exact reasoning?

Yeah there could be parts of the training set with 1611 being explicitly called the official and 1622 being explicitly called the most common answer. But it can also have access to search results directly I think. Is there a way to know if it does or not?